1. Field of the Invention
This invention pertains to the field of targeting of gene delivery systems (viral and non-viral) to particular cell and tissue types.
2. Background
The use of recombinant viral vectors for the delivery of exogenous genes to mammalian cells is well established. See e.g. Boulikas, T. in Gene Therapy and Molecular Biology Volume 1, pages 1-172 (Boulikas, Ed.) 1998, Gene Therapy Press, Palo Alto, Calif. However, certain viral vectors commonly used in such instances, such as adenoviruses, exhibit a broad tropism which permits infection and expression of the exogenous gene in a variety of cell types. While this can be useful in some instances, the treatment of certain diseases is enhanced if the virus is able to be modified so as xe2x80x9ctargetxe2x80x9d (i.e., to preferentially infect) only a limited type of cell or tissue.
A variety of approaches to create targeted viruses have been described in the literature. For example, cell targeting has been achieved with adenovirus vectors by selective modification of the viral genome knob and fiber coding sequences to achieve expression of modified knob and fiber domains having specific interaction with unique cell surface receptors. Examples of such modifications are described in Wickham et al. (1997) J. Virol. 71(11):8221-8229 (incorporation of RGD peptides into adenoviral fiber proteins); Arnberg et al. (1997) Virology 227:239-244 (modification of adenoviral fiber genes to achieve tropism to the eye and genital tract); Harris and Lemoine (1996) TIG 12(10):400-405; Stevenson et al. (1997) J. Virol. 71(6):4782-4790; Michael et al. (1995) Gene Therapy 2:660-668 (incorporation of gastrin releasing peptide fragment into adenovirus fiber protein); and Ohno et al. (1997) Nature Biotechnology 15:763-767 (incorporation of Protein A-IgG binding domain into Sindbis virus).
However, the design of a functional chimeric protein for targeting is not facile. For example, if one wishes to create a chimeric adenoviral knob protein containing an targeting domain, the recombinant knob protein must be able to (a) assemble properly into the icosahedral viral structure and (b) also retain the binding specificity of the targeting moiety. This may involve significant and complex molecular modeling to incorporate the targeting moiety into the appropriate region of the knob protein to insure that the targeting moiety is on the surface of the knob protein. Additionally, since the precise process for assembly of the adenoviral particle is poorly understood it is possible that insertion of a large targeting moiety will sufficiently interrupt the three dimensional structure of the viral protein so that it does not efficiently assemble into an infectious virion. Furthermore, whenever one wishes to change the targeting properties of the adenovirus, it is necessary to reengineer the knob protein taking into account all of the foregoing, which can be a lengthy process. Moreover, the manipulation of the adenoviral genome to obtain a gene that encodes the chimeric protein is a time consuming process, due to the size and complexity of the adenoviral genome.
In order to avoid these hurdles, other methods of cell specific targeting rely on the conjugation of antibodies or antibody fragments to the envelope proteins (see, e.g. Michael et al. (1993) J. Biol. Chem. 268:6866-6869, Watkins et al. (1997) Gene Therapy 4:1004-1012; Douglas et al. (1996) Nature Biotechnology 14: 1574-1578. This approach also has its limitations. First, in the case of chemically conjugating the antibody (or antibody fragment) to the surface of the virion, the linkage is generally achieved by modification of amino acyl side chains in the antibody (particularly through lysine residues). As it is difficult to control the stoichiometry of this reaction, one can envision the resulting virion being coated with antibodies in a variety of orientations. As the binding specificity of the antibody is contained in the variable regions, the random association of the cross-linked antibody will result in many of the antibody variable domains being xe2x80x9chiddenxe2x80x9d and thus ineffective. Accordingly, in order to insure a sufficient number of exposed variable domains to achieve efficient targeting, a significant excess of antibody must be complexed to the virion. Additionally, the coating of the virion with an excess of antibodies may interfere with internalization of the virus in the target cell. For example, in the case of adenoviruses, the interaction between the viral coat proteins and the CAR receptor is believed to be an essential step in the infectious process. If the viral coat proteins are obscured by an excess of antibody proteins, one may expect that the efficiency of binding to the CAR receptor and internalization would suffer. If the virion is unable to infect the cell and exert its therapeutic effect, it is questionable whether this targeting approach would provide significant therapeutic benefit.
Alternative to the use of antibodies, others have complexed targeting proteins to the surface of the virion. See, e.g. Nilson et al. (1996) Gene Therapy 3:280-286 (conjugation of EGF to retroviral proteins). However, this approach suffers many of the same limitation as the use of antibodies, such as obscuring viral coat proteins and potentially interfering with the infectious mechanism.
In one attempt to avoid these problems, some groups have used anti-knob or anti-fiber antibodies complexed to a targeting moiety (see, e.g., U.S. Pat. No. 5,871,727). While this avoids the problem of having a antibody-coated virion as discussed above, such non-covalent complexes are in equilibrium with the free conjugated antibody and virion species, i.e.
{conjugated antibody-virion}⇄conjugated-antibody+virion.
While the affinity of the antibody for the knob may be high and the resulting equilibrium constant of this reaction suggests the formation of a xe2x80x9cstablexe2x80x9d complex, this does not indicate that the complex will be kinetically stable in solution over a period of time. Additionally, although a complex may be xe2x80x9cstablexe2x80x9d in a solution of limited volume, upon introduction of the complex to a solution of essentially infinite volume (e.g., the bloodstream of a mammal) the equilibrium will be shifted in favor of dissociation of such a complex.
The present invention provides targeted complexes that are useful for delivering molecules to a particular cell or tissue type of interest. The invention provides targeted complexes of the formula:
{(delivery vehicle-CM)-TMI-(CM-targeting ligand)};
The delivery vehicle can be, for example, a peptide vector, a peptide-DNA aggregate, a liposome, a gas-filled microsome, an encapsulated macromolecule, and the like. In some embodiments, the delivery vehicle is a viral vector. Particularly suitable viral vectors include a retrovirus, a vaccinia virus, a herpes virus, an adeno-associated virus, a minute virus of mice (MVM), a human immunodeficiency virus, a sindbis virus, an MoMLV, and a hepatitis virus.
xe2x80x9cCMxe2x80x9d is a chelating moiety, such as a chelating peptide or an organic chelating agent. TMI is a transition metal ion. CM-targeting ligand is a chelating moiety (CM) covalently linked to a targeting ligand that can bind to a cell or tissue of interest.
The invention also provides methods for producing a kinetically inert targeted delivery vehicle complex. These methods involve: a) preparing a kinetically labile transition metal complex by contacting a delivery vehicle-CM and a CM-targeting ligand with a transition metal ion that is in a kinetically labile oxidation state; and b) changing the oxidation state of the metal ion to form the kinetically inert complex.
Also provided by the invention are methods of delivering a therapeutic or diagnostic agent to a target cell in an organism. These methods involve administering to an organism a targeted complex of the formula:
{(delivery vehicle-CM)-TMI-(CM-targeting ligand)};
wherein delivery vehicle-CM is a delivery vehicle that displays on its surface a polypeptide that comprises a chelating moiety (CM), TMI is a transition metal ion, and CM-targeting ligand is a chelating moiety (CM) covalently linked to a targeting ligand that binds to the target cell.